Datasets:
| { | |
| "batch_size": 2, | |
| "checkpoint": "main", | |
| "correctness_split": false, | |
| "layers": [ | |
| 21 | |
| ], | |
| "model_id": "LLM360/Amber", | |
| "num_samples": 80, | |
| "representation_type": "residual_stream", | |
| "seed": 42, | |
| "split_stats": { | |
| "all": { | |
| "n_examples": 80, | |
| "skipped": false | |
| } | |
| }, | |
| "task": "compositional:lower_first", | |
| "vector_norms": { | |
| "all": { | |
| "21": 0.9999999403953552 | |
| } | |
| } | |
| } | |